Azure Blob Storage Error/Failure Rates...



  • Hi folks,

    I am trying to find data on Azure blob failure rates. Ideally, previous downtime/unavailability numbers, error/corruptuon rates, total loss rates... but I cannot find any data on this. Can anybody point me in the right direction?

    Perhaps I'm wrong and it just 'never fails', but I can't believe that. If a blob storage can never fail, then why sell the option of locally redundant sync between Azure data centres and global redundance sync to different countries...

    Best,
    Jim



  • @Jimmy9008 said in Azure Blob Storage Error/Failure Rates...:

    If a blob storage can never fail, then why sell the option of locally redundant sync between Azure data centres and global redundance sync to different countries...

    Because the connection to the blob can go down. This is the same as AWS selling service plans to have your data in multiple data centers. Look at the freak out from the February 2017 AWS outage on the East coast.

    The only people that were "down" were people that chose not to pay to have their stuff in more than one availability zone.



  • @Jimmy9008 said in Azure Blob Storage Error/Failure Rates...:

    Perhaps I'm wrong and it just 'never fails', but I can't believe that.

    Define fail? An Azure Platform error that corrupts it? I'm sure it is possible, but I've never heard of it. I've also never went looking for it.

    By far the larger problem will always be your application screwing up something inside the blob storage. That is not Azure's problem to worry about.



  • I could not find anything specific to blob storage, but I would imagine it's setup like Amazon s3 with 99.999999999% availability. Since AWS and Azure seem to walk step by step, I'm guessing it's very similar since they are competing products.

    S3 by default is across all availability zones in a region. So there would have to be catastrophic region failure and multiple facilities would have gone down.

    I know with both Azure and AWS you can have object storage in multiple regions and deliver it through their CDN service (Azure CDN or AWS Cloudfront



  • Here is some information about s3. I found out most of the time both aws and azure are setup the same with competing services.

    https://aws.amazon.com/s3/faqs/



  • @Jimmy9008 said in Azure Blob Storage Error/Failure Rates...:

    If a blob storage can never fail, then why sell the option of locally redundant sync between Azure data centres and global redundance sync to different countries...

    Generally when you spread out cross region, it's for performance not necessarily availability (well at least ast in the case of object storage).



  • @JaredBusch said in Azure Blob Storage Error/Failure Rates...:

    Define fail? An Azure Platform error that corrupts it? I'm sure it is possible, but I've never heard of

    Right. Durability is likely at ten nines or higher. The data isn't going anywhere.

    If Availability is the concern, chances are those failures will be on your end not Azure's, but Azure definitely has outages (as does everyone else), but you can mitigate this with regional redundancy. Azure itself does have higher than normal outages from what we've seen, but like anything, Azure itself isn't a solution, just a building block for your solution. So the resulting availability is more of what you make out of it than was MS provides directly.




Log in to reply